Language is a Virus: Locales and PostgreSQL

Language is a Virus: Locales and PostgreSQL

This is a pragmatic look at PostgreSQL's support for locales, in particular around character encoding and collations. This is intended for DBAs and developers who want to make use of PostgreSQL's extensive locale features with a minimum number of gotchas. Among the topics are:

* A quick overview of locales, character encodings, and collations.

* What a "locale provider," and how do you use it?

* How do various collations perform, really?

* What are the use cases for various locales?

* And what is this horrible thing about index corruption I hear about?

The goal is to provide developers and DBAs enough information to make informed decisions about how to apply the PostgreSQL locale system in a real-life database, and how to avoid pitfalls and gotchas.